.\" $Id$
.\"
.\" Copyright (c) 2009 Kristaps Dzonsons <kristaps@kth.se>
.\"
.\" Permission to use, copy, modify, and distribute this software for any
.\" purpose with or without fee is hereby granted, provided that the
.\" above copyright notice and this permission notice appear in all
.\" copies.
.\"
.\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL
.\" WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED
.\" WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE
.\" AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
.\" DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
.\" PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
.\" TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
.\" PERFORMANCE OF THIS SOFTWARE.
.\"
.Dd $Mdocdate$
.Dt mdoc 3
.Os
.\" SECTION
.Sh NAME
.Nm mdoc_alloc ,
.Nm mdoc_parseln ,
.Nm mdoc_endparse ,
.Nm mdoc_node ,
.Nm mdoc_meta ,
.Nm mdoc_free
.Nd mdoc macro compiler library
.\" SECTION
.Sh SYNOPSIS
.Fd #include <mdoc.h>
.Vt extern const char * const * mdoc_macronames;
.Vt extern const char * const * mdoc_argnames;
.Ft "struct mdoc *"
.Fn mdoc_alloc "void *data" "const struct mdoc_cb *cb"
.Ft void
.Fn mdoc_free "struct mdoc *mdoc"
.Ft int
.Fn mdoc_parseln "struct mdoc *mdoc" "int line" "char *buf"
.Ft "const struct mdoc_node *"
.Fn mdoc_node "struct mdoc *mdoc"
.Ft "const struct mdoc_meta *"
.Fn mdoc_meta "struct mdoc *mdoc"
.Ft int
.Fn mdoc_endparse "struct mdoc *mdoc"
.\" SECTION
.Sh DESCRIPTION
The
.Nm mdoc
library parses lines of mdoc input into an abstract syntax tree.
.Dq mdoc ,
which is used to format BSD manual pages, is a macro package of the
.Dq roff
language. The
.Nm
library implements only those macros documented in the
.Xr mdoc 7
and
.Xr mdoc.samples 7
manuals. Documents with
.Xr refer 1 ,
.Xr eqn 1
and other pre-processor sections aren't accomodated.
.\" PARAGRAPH
.Pp
.Nm
is
.Ud
.\" PARAGRAPH
.Pp
In general, applications initiate a parsing sequence with
.Fn mdoc_alloc ,
parse each line in a document with
.Fn mdoc_parseln ,
close the parsing session with
.Fn mdoc_endparse ,
operate over the syntax tree returned by
.Fn mdoc_node
and
.Fn mdoc_meta ,
then free all allocated memory with
.Fn mdoc_free .
See the
.Sx EXAMPLES
section for a full example.
.\" PARAGRAPH
.Pp
This section further defines the
.Sx Types ,
.Sx Functions
and
.Sx Variables
available to programmers. Following that,
.Sx Character Encoding
describes input format. Lastly,
.Sx Abstract Syntax Tree ,
documents the output tree.
.\" SUBSECTION
.Ss Types
Both functions (see
.Sx Functions )
and variables (see
.Sx Variables )
may use the following types:
.Bl -ohang -offset "XXXX"
.\" LIST-ITEM
.It Vt struct mdoc
An opaque type defined in
.Pa mdoc.c .
Its values are only used privately within the library.
.\" LIST-ITEM
.It Vt struct mdoc_cb
A set of message callbacks defined in
.Pa mdoc.h .
.\" LIST-ITEM
.It Vt struct mdoc_node
A parsed node. Defined in
.Pa mdoc.h .
See
.Sx Abstract Syntax Tree
for details.
.El
.\" SUBSECTION
.Ss Functions
Function descriptions follow:
.Bl -ohang -offset "XXXX"
.\" LIST-ITEM
.It Fn mdoc_alloc
Allocates a parsing structure. The
.Fa data
pointer is passed to callbacks in
.Fa cb ,
which are documented further in the header file. Returns NULL on
failure. If non-NULL, the pointer must be freed with
.Fn mdoc_free .
.\" LIST-ITEM
.It Fn mdoc_free
Free all resources of a parser. The pointer is no longer valid after
invocation.
.\" LIST-ITEM
.It Fn mdoc_parseln
Parse a nil-terminated line of input. This line should not contain the
trailing newline. Returns 0 on failure, 1 on success. The input buffer
.Fa buf
is modified by this function.
.\" LIST-ITEM
.It Fn mdoc_endparse
Signals that the parse is complete. Note that if
.Fn mdoc_endparse
is called subsequent to
.Fn mdoc_node ,
the resulting tree is incomplete. Returns 0 on failure, 1 on success.
.\" LIST-ITEM
.It Fn mdoc_node
Returns the first node of the parse. Note that if
.Fn mdoc_parseln
or
.Fn mdoc_endparse
return 0, the tree will be incomplete.
.It Fn mdoc_meta
Returns the document's parsed meta-data. If this information has not
yet been supplied or
.Fn mdoc_parseln
or
.Fn mdoc_endparse
return 0, the data will be incomplete.
.El
.\" SUBSECTION
.Ss Variables
The following variables are also defined:
.Bl -ohang -offset "XXXX"
.\" LIST-ITEM
.It Va mdoc_macronames
An array of string-ified token names.
.\" LIST-ITEM
.It Va mdoc_argnames
An array of string-ified token argument names.
.El
.\" SUBSECTION
.Ss Character Encoding
The
.Xr mdoc 3
library accepts only printable ASCII characters as defined by
.Xr isprint 3 .
Non-ASCII character sequences are delimited in various ways. All are
preceeded by an escape character
.Sq \\
and followed by either an open-parenthesis
.Sq \&(
for two-character sequences; an open-bracket
.Sq \&[
for n-character sequences (terminated at a close-bracket
.Sq \&] ) ;
an asterisk and open-parenthesis
.Sq \&*(
for two-character sequences;
an asterisk and non-open-parenthesis
.Sq \&*
for single-character sequences; or one of a small set of standalone
single characters for other escapes.
.\" PARAGRAPH
.Pp
Examples:
.Pp
.Bl -tag -width "XXXXXXXX" -offset "XXXX" -compact
.\" LIST-ITEM
.It \\*(<=
prints
.Dq \*(<=
.Pq greater-equal
.\" LIST-ITEM
.It \\(<-
prints
.Dq \(<-
.Pq left-arrow
.\" LIST-ITEM
.It \\[<-]
also prints
.Dq \(<-
.Pq left-arrow
.\" LIST-ITEM
.It \\*(Ba
prints
.Dq \*(Ba
.Pq bar
.\" LIST-ITEM
.It \\*q
prints
.Dq \*q
.Pq double-quote
.El
.\" PARAGRAPH
.Pp
All escaped sequences are syntax-checked, but it's up to the front-end
system to correctly render them to the output device.
.\" SUBSECTION
.Ss Abstract Syntax Tree
The
.Nm
functions produce an abstract syntax tree (AST) describing the input
lines in a regular form. It may be reviewed at any time with
.Fn mdoc_nodes ;
however, if called before
.Fn mdoc_endparse ,
or after
.Fn mdoc_endparse
or
.Fn mdoc_parseln
fail, it may be incomplete.
.\" PARAGRAPH
.Pp
The AST is composed of
.Vt struct mdoc_node
nodes with block, head, body, element, root and text types as declared
by the
.Va type
field. Each node also provides its parse point (the
.Va line ,
.Va sec ,
and
.Va pos
fields), its position in the tree (the
.Va parent ,
.Va child ,
.Va next
and
.Va prev
fields) and type-specific data (the
.Va data
field).
.\" PARAGRAPH
.Pp
The tree itself is arranged according to the following normal form,
where capitalised non-terminals represent nodes.
.Pp
.Bl -tag -width "ELEMENTXX" -compact -offset "XXXX"
.\" LIST-ITEM
.It ROOT
\(<- mnode+
.It mnode
\(<- BLOCK | ELEMENT | TEXT
.It BLOCK
\(<- (HEAD [TEXT])+ [BODY [TEXT]] [TAIL [TEXT]]
.It BLOCK
\(<- BODY [TEXT] [TAIL [TEXT]]
.It ELEMENT
\(<- TEXT*
.It HEAD
\(<- mnode+
.It BODY
\(<- mnode+
.It TAIL
\(<- mnode+
.It TEXT
\(<- [[:alpha:]]*
.El
.\" PARAGRAPH
.Pp
Of note are the TEXT nodes following the HEAD, BODY and TAIL nodes of
the BLOCK production. These refer to punctuation marks. Furthermore,
although a TEXT node will generally have a non-zero-length string, in
the specific case of
.Sq \&.Bd \-literal ,
an empty line will produce a zero-length string.
.\" PARAGRAPH
.Pp
The rule-of-thumb for mapping node types to macros follows. In-line
elements, such as
.Sq \&.Em foo ,
are classified as ELEMENT nodes, which can only contain text.
Multi-line elements, such as
.Sq \&.Sh ,
are BLOCK elements, where the HEAD constitutes line contents and the
BODY constitutes subsequent lines. In-line elements with matching
pairs, such as
.Sq \&.So
and
.Sq \&.Sc ,
are BLOCK elements with no HEAD tag. The only exception to this is
.Sq \&.Eo
and
.Sq \&.Ec ,
which has a HEAD and TAIL node corresponding to the enclosure string.
TEXT nodes, obviously, constitute text, and the ROOT node is the
document's root.
.\" SECTION
.Sh EXAMPLES
The following example reads lines from stdin and parses them, operating
on the finished parse tree with
.Fn parsed .
Note that, if the last line of the file isn't newline-terminated, this
will truncate the file's last character (see
.Xr fgetln 3 ) .
Further, this example does not error-check nor free memory upon failure.
.Bd -literal -offset "XXXX"
struct mdoc *mdoc;
struct mdoc_node *node;
char *buf;
size_t len;
int line;
line = 1;
mdoc = mdoc_alloc(NULL, NULL);
while ((buf = fgetln(fp, &len))) {
buf[len - 1] = '\\0';
if ( ! mdoc_parseln(mdoc, line, buf))
errx(1, "mdoc_parseln");
line++;
}
if ( ! mdoc_endparse(mdoc))
errx(1, "mdoc_endparse");
if (NULL == (node = mdoc_node(mdoc)))
errx(1, "mdoc_node");
parsed(mdoc, node);
mdoc_free(mdoc);
.Ed
.\" SECTION
.Sh SEE ALSO
.Xr mdoc 7 ,
.Xr mdoc.samples 7 ,
.Xr groff 1 ,
.Xr mdocml 1
.\" SECTION
.Sh AUTHORS
The
.Nm
utility was written by
.An Kristaps Dzonsons Aq kristaps@kth.se .
.\" SECTION
.Sh BUGS
Bugs, un-implemented macros and incompabilities are documented in this
section. The baseline for determining whether macro parsing is
.Qq incompatible
is the default
.Xr groff 1
system bundled with
.Ox .
.\" PARAGRAPH
.Pp
Un-implemented: the
.Sq \&Xc
and
.Sq \&Xo
macros aren't handled when used to span lines for the
.Sq \&It
macro. Such usage is specifically discouraged in
.Xr mdoc.samples 7 .
.\" PARAGRAPH
.Pp
Bugs: when
.Sq \&It \-column
is invoked, whitespace is not stripped around
.Sq \&Ta
or tab-character separators.
.\" PARAGRAPH
.Pp
Bugs: elements within columns for
.Sq \&It \-column
are not yet supported.
.\" PARAGRAPH
.Pp
Incompatible: the
.Sq \&At
macro only accepts a single parameter. Furthermore, several macros
.Pf ( Sq \&Pp ,
.Sq \&It ,
and possibly others) accept multiple arguments with a warning.
.\" PARAGRAPH
.Pp
Incompatible: only those macros specified by
.Xr mdoc.samples 7
and
.Xr mdoc 7
for
.Ox
are supported; support for
.Nx
and other
.Bx
systems is in progress.
